On a related topic. Have you seen:http://www.seedmagazine.com/news/2007/05/dirty_little_secret.phpMost of the related links are in the discussion:http://skepchick.org/skepticsguide/viewtopic.php?t=3571&start=0The paper by Sterne and Smith is also easy to find. I am hoping that there will be a comprehensive discussion on this topic but my background in Statistics is not strong enough for a good discussion.
GS - very interesting. I think, in clinical stuff, a lot of the erroneous conclusions are due to poor understanding of statistics. Not all the wrong results may necessarily be due to conscious fraud; but fraud, too, is more widespread than we like to acknowledge. And if carefully and cleverly done, it is almost undetectable.
Post a Comment